CHAPTER 1
INTRODUCTION
1.1
Background of the problem
For the last ten years, business and manufacturing have shifted from local entities
towards globalization entities. Quality has become one of the most important decision
factors in making selection among worldwide competing products and service. To be a
winner in this global competitive edge, the manufacturer needs to produce product with
high quality within the shortest delivery time and lowest cost (Shariff, 2005).
There are several methods which can be used to improve product quality. One of
the commonly used methods is the Statistical Process Control (SPC). SPC can be
defined as using statistical techniques to improve and implement quality control. Since
the introduction of SPC by Walter A Shewart in year 1920s, SPC have been widely
adopted and applied in manufacturing industrial. It has been used for monitoring and
diagnosis manufacturing process variation. SPC can be used to achieve process stability
and improve process capability through the reduction of process variability.
Currently there are several computer aided SPC systems introduced by software
vendors and researchers. These computer aided SPC systems can assist the
manufacturers in automating the quality control functions. Nevertheless, these systems
2
still have some limitations. For example, the current computer aided SPC system only
provides some basic functions such as automating the plotting of control charts. The
most important part, the control chart data analysis still heavily relies on the operators
and engineers. Usually, the operator and engineer are very busy with other production
task and may not have sufficient time to perform analysis on the control chart data
(Thompson et al., 1998).
As noted earlier, the current computer aided SPC system are mainly based on
standalone system. These standalone systems do not allow co-operation and information
sharing. The accessibility of these control chart data should not be limited to the operator
and engineer only. They should also be accessible to managers and customers. The
control chart data can assist the manager or customer to have a better understanding on
production line status and helps them in making better decision.
Beside that, the current computer aided SPC system does not allow remote
access. It is only limited to one production area or factory side. The operator and
engineer cannot access the status of production line remotely. Some managers and
customers may not have quality engineering background. The burden of analysis the
production data should not be assigned to them. The manager and customer maybe
interested to know the result and status of the production line. An automated and
computerized SPC system should assist them to diagnose the production line.
These are the quality control problems facing by the operator, engineer, manager
and customer. So there is a need to develop a Client-Server based Control Chart Pattern
Recognition (CCPR) system which would enable data sharing, remote access and
intelligent diagnosis of the production line status.
3
1.2
Statement of the problem
Existing control chart pattern recognition systems are mainly local and
standalone system. They do not provided data sharing and remote access. There is a
need to enhance the standalone system towards a Client-server based control chart
patterns recognition system. Several design issues need to be considered in developing
such a client-server system.
The client-server based CCPR system needs to have a client-server architecture
which can support multiple users simultaneously and better resource allocation.
Inappropriate allocation of resources can cause certain functional tier become
overloaded while the others are under loaded. An effective assignment of functional
modules can enable the CCPR system to support multiple users simultaneously without
additional computing resources.
Due to unique characteristics of SPC data stream, the load passing procedure as
implemented in the voice and network packet system are not directly applicable in the
CCPR system. A customized load passing procedure is needed for a client-server based
CCPR system. Specifically, the load passing algorithm must be able to prioritize the
client access based on the severity of process instability. Without a customized load
passing procedure, the unstable process will have the same priority with stable process,
which may result in delayed detection of unstable process data.
4
1.3
Purpose and Objective of the Study
The purpose of this study is to enhance the existing standalone control chart
patterns recognition system towards a client-server based system which would enable
efficient data sharing, remote access and intelligent diagnosis of production line.
The specific objectives of this study are:
i.
To design a customized client and server architecture for a control chart patterns
recognition system.
ii.
To customize a load passing procedure for the client-server based control chart
patterns recognition system.
iii.
To provide a guideline for transforming the standalone system to a client-server
based CCPR system.
1.4
Scope of the Study
Below are the scopes for this study:
i.
Data stream for process variation are limited to univariate data plotted of Shewart
x -chart.
ii.
Design and evaluation are limited to the computer simulation studies. The
published and real data are used for validation. .
iii.
The simulation data arrive to the system with constant rate and speed.
iv.
Different combination of computing resources are not taken into consideration.
Fixed type and amount of computing resources are used.
v.
Data encryption and network security issues are not included in this study.
5
1.5
Significance of the study
This client-server based CCPR system enables information sharing among
different entities in the manufacturing system. By taking the advantageous of internet
technology, this web-enabled system is accessible anytime, anywhere using a personal
computer with Internet connection. Client-server based CCPR system makes the
manager and customer feel easy and comfortable while interacting with the control chart
system. The suitable client and server architecture ensures each functional tier can be
successfully integrated. This architecture should be able to support multiple users
simultaneously and ensure fair distribution of the system workload. With all these
features, this client-server based control chart patterns recognition system could
overcome the limitation of the standalone system. This study also provides a guideline
for transformation from a standalone to the client-server based CCPR system.
1.6
Organization of the Thesis
This report is organized into seven chapters. Chapter 1 gives explanation on the
background, statement of problem, objective, scope and significance of study. Chapter 2
presents literature review on board and focus area of the study. Techniques and
technologies for the client-server based system development are presented here. Chapter
3 discusses the research methodology for this study. It begins with discussion on the
problem situation, solution concept followed by the researches planning, source of data,
equipment used and performance measurement methods. Chapter 4 describes the
experiments and statistical analysis carried out to select a suitable client and server
architecture. Chapter 5 describes the experiment and statistical analysis carried out for
selecting a customized load passing procedure for the client-server based CCPR system.
Chapter 6 presents the techniques and methods used for the prototype system
implementation and validation. Lastly, Chapter 7 summarizes the research finding and
concludes the study.